Use of medication data to validate an association in community-based symptom prevalence studies

Hari H. Dayal, Yi Hwei Li, Wayne Snodgrass, Vivek Dayal, Chandra K. Mittal

Research output: Contribution to journalArticlepeer-review

10 Scopus citations


A chemical spill from an oil refinery in Texas City, Texas, exposed the community to more than 40 000 lbs (18 144 kg) of highly toxic and corrosive hydrofluoric acid. A symptom prevalence study indicated an association between symptom reports, most notably breathing symptoms, and hydrofluoric acid exposure. Although verification of self-reported symptoms by checking medical records or performing clinical tests is theoretically possible, it is not a feasible alternative in dealing with an entire community. Open-ended data on medication use collected in the prevalence study were coded by organ system and analyzed by cross-classification techniques and log linear models. Results showed that the reported use of medication for hydrofluoric acid-related problems was associated with the exposure; medication use for problems unrelated to hydrofluoric acid exposure was uniform across the exposure categories. Moreover, medication use was significantly associated with the severity of breathing-related problems for each exposure category. Medication use, however, may have been under-reported because it seems difficult to conjure up the names of medications that were not taken or medications not taken recently may not be recalled. Nonetheless, open-ended medication data may be a useful surrogate approach to validating an association between an exposure and health outcomes.

Original languageEnglish (US)
Pages (from-to)93-97
Number of pages5
JournalArchives of Environmental Health
Issue number2
StatePublished - 1994

ASJC Scopus subject areas

  • Environmental Chemistry
  • General Environmental Science
  • Public Health, Environmental and Occupational Health
  • Health, Toxicology and Mutagenesis


Dive into the research topics of 'Use of medication data to validate an association in community-based symptom prevalence studies'. Together they form a unique fingerprint.

Cite this